Electric Power Systems Research 78 (2008) 19891996

Contents lists available at ScienceDirect

Electric Power Systems Research

journal homepage: www.elsevier.com/locate/epsr

Review

A summary of demand response in electricity markets

M.H. Albadi , E.F. El-Saadany
Department of Electrical and Computer Engineering, University of Waterloo, 200 University Ave. W, Waterloo, ON N2L3G1, Canada

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents a summary of Demand Response (DR) in deregulated electricity markets. The def-
Received 19 October 2007 inition and the classication of DR as well as potential benets and associated cost components are
Received in revised form 7 April 2008 presented. In addition, the most common indices used for DR measurement and evaluation are high-
Accepted 8 April 2008
lighted, and some utilities experiences with different demand response programs are discussed. Finally,
Available online 21 May 2008
the effect of demand response in electricity prices is highlighted using a simulated case study.
2008 Elsevier B.V. All rights reserved.
Keywords:
Demand response
Price elasticity
Real time pricing
Electricity markets

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1989
2. Denition and classication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1990
2.1. Denition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1990
2.2. Customer response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1990
2.3. Program classication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1990
3. DR benets and costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1991
3.1. DR benets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1991
3.2. DR costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1992
4. DR measurement and simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1992
4.1. DR measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1992
4.2. Market simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1993
4.3. Optimal Power Flow formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1993
4.4. Simulation results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1993
4.5. DR experiences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1994
5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1995
Appendix A. List of symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1995
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1995

1. Introduction ditional approach was to supply all power demands whenever they
occurred, however, the new philosophy states that the system will
For many reasons, electric utilities and power network compa- be most efcient if uctuations in demand are kept as small as
nies have been forced to restructure their operations from vertically possible.
integrated mechanisms to open market systems [1]. With the Reliable operation of the electricity system necessitates a per-
restructuring and deregulation of the electricity supply industry, fect balance between supply and demand in real time. This balance
the philosophy of operating the system was also changed. The tra- is not easy to achieve given that both supply and demand levels
can change rapidly and unexpectedly due to many reasons, such as
generation unit forced outages, transmission and distribution line
Corresponding author. Tel.: +1 519 888 4567x37059; fax: +1 519 746 3077. outages, and sudden load changes. The electricity system infras-
E-mail address: mhhalbad@uwaterloo.ca (M.H. Albadi). tructure is highly capital intensive; demand side (load) response

0378-7796/$ see front matter 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsr.2008.04.002
1990 M.H. Albadi, E.F. El-Saadany / Electric Power Systems Research 78 (2008) 19891996
is one of the cheaper resources available for operating the system
according to the new philosophy.
This paper presents an overview of new exible resources for
operating a reliable system. The paper starts with dening the
Demand Response (DR) and how electricity consumers can be
responsive. Highlighting different DR programs follows, includ-
ing classical, new market-based and dynamic pricing scenarios.
Potential cost savings and benets related to different market com-
ponents are also discussed. Measuring indices are used to assess
program success. The paper is concluded with a few selected utili-
ties experience with DR programs.
2. Denition and classication
2.1. Denition
Demand response can be dened as the changes in electricity
usage by end-use customers from their normal consumption pat-
terns in response to changes in the price of electricity over time.
Further, DR can be also dened as the incentive payments designed
to induce lower electricity use at times of high wholesale market
prices or when system reliability is jeopardized [2]. DR includes
all intentional electricity consumption pattern modications by
end-use customers that are intended to alter the timing, level of
instantaneous demand, or total electricity consumption [3].
2.2. Customer response
There are three general actions by which a customer response
can be achieved [2]. Each of these actions involves cost and mea-
sures taken by the customer. First, customers can reduce their
electricity usage during critical peak periods when prices are high
without changing their consumption pattern during other periods.
This option involves a temporary loss of comfort. This response
is achieved, for instance, when thermostat settings of heaters
or air conditioners are temporary changed [4,5]. Secondly, cus-
tomers may respond to high electricity prices by shifting some of
their peak demand operations to off-peak periods, as an exam-
ple, they shift some household activities (e.g., dishwashers, pool
pumps) to off-peak periods. The residential customer in this case
will bear no loss and will incur no cost. However, this will not
be the case if an industrial customer decides to reschedule some
activities and rescheduling costs to make up for lost services are
incurred. The third type of customer response is by using onsite
generationcustomer owned distributed generation [6,7]. Cus-
tomers who generate their own power may experience no or very
little change in their electricity usage pattern; however, from utility
prospective, electricity use patterns will change signicantly, and
demand will appear to be smaller.
2.3. Program classication
Different DR programs are shown in Fig. 1. These programs can
be classied into two main categories: Incentive-Based Programs
(IBP) and Price-Based Programs (PBP) [2,8]. Some literature papers
named these categories as system- and market-led, emergency-
and economic-based, or stability- and economic-based DR pro-
grams [3,9,10]. IBP are further divided into classical programs and
market-based programs. Classical IBP include Direct Load Control
programs and Interruptible/Curtailable Load programs. Market-
based IBP include Emergency DR Programs, Demand Bidding,
Capacity Market, and the Ancillary services market. In classical IBP,
participating customers receive participation payments, usually as
a bill credit or discount rate, for their participation in the programs.
In market-based programs, participants are rewarded with money
Fig. 1. Classication of demand response programs.
for their performance, depending on the amount of load reduction
during critical conditions.
In Direct Load Control programs, utilities have the ability to
remotely shut down participant equipment on a short notice. Typ-
ical remotely controlled equipment includes air conditioners and
water heaters. This kind of programs is of interest mainly to resi-
dential customers and small commercial customers.
As with Direct Load Control programs, customers participat-
ing in Interruptible/Curtailable Programs receive upfront incentive
payments or rate discounts. Participants are asked to reduce their
load to predened values. Participants who do not respond can face
penalties, depending on the program terms and conditions.
Demand Bidding (also called Buyback) programs are programs
in which consumers bid on specic load reductions in the electricity
wholesale market. A bid is accepted if it is less than the market price.
When a bid is accepted, the customer must curtail his load by the
amount specied in the bid or face penalties. On the other hand, in
Emergency DR Programs, participating customers are paid incen-
tives for measured load reductions during emergency conditions
[2].
Furthermore, Capacity Market Programs are offered to cus-
tomers who can commit to providing pre-specied load reductions
when system contingencies arise [2]. Participants usually receive a
day-ahead notice of events and are penalized if they do not respond
to calls for load reduction. Ancillary services market programs allow
customers to bid on load curtailment in the spot market as operat-
ing reserve. When bids are accepted, participants are paid the spot
market price for committing to be on standby and are paid spot
market energy price if load curtailment is required [2].
PBP programs are based on dynamic pricing rates in which elec-
tricity tariffs are not at; the rates uctuate following the real time
cost of electricity. The ultimate objective of these programs is to
atten the demand curve by offering a high price during peak peri-
ods and lower prices during off-peak periods. These rates include
the Time of Use (TOU) rate, Critical Peak Pricing (CPP), Extreme Day
Pricing (EDP), Extreme Day CPP (ED-CPP), and Real Time Pricing
(RTP). The basic type of PBP is the TOU rates, which are the rates of
electricity price per unit consumption that differ in different blocks
of time. The rate during peak periods is higher than the rate during
off-peak periods. The simplest TOU rate has two time blocks; the
peak and the off-peak. The rate design attempts to reect the aver-
age cost of electricity during different periods. A TOU rate design
 M.H. Albadi, E.F. El-Saadany / Electric Power Systems Research 78 (2008) 19891996
Fig. 2. Benets associated with DR.
process is described in [11]. CPP rates include a pre-specied higher
electricity usage price superimposed on TOU rates or normal at
rates. CPP prices are used during contingencies or high wholesale
electricity prices for a limited number of days or hours per year
[2,10]. On the other hand, EDP is similar to CPP in having a higher
price for electricity and differs from CPP in the fact that the price is
in effect for the whole 24 h of the extreme day, which is unknown
until a day-ahead [12]. Furthermore, in ED-CPP rates, CPP rates for
peak and off-peak periods are called during extreme days. How-
ever, a at rate is used for the other days [12]. RTP are programs in
which customers are charged hourly uctuating prices reecting
the real cost of electricity in the wholesale market. RTP customers
are informed about the prices on a day-ahead or hour-ahead basis.
Many economists are convinced that RTP programs are the most
direct and efcient DR programs suitable for competitive electricity
markets and should be the focus of policymakers [13].
3. DR benets and costs
This section covers and discusses both potential benets
expected from DR programs and the associated cost.
3.1. DR benets
Fig. 2 summarizes the benets associated with DR; they fall
under four main categories: participant, market-wide, reliability,
and market performance benets.
Customers participating in DR programs can expect savings in
electricity bills if they reduce their electricity usage during peak
periods [2,8,10,14]. In fact, some participants may experience sav-
ings even if they do not change their consumption pattern if their
normal consumption during high price peak periods is lower than
their class average [12]. Some customers might be able to increase
their total energy consumption without having to pay more money
by operating more off-peak equipment. Moreover, participants in
classical IBP are entitled to receive incentive payments for their
participation, while market-based IBP customers will receive pay-
ments according to their performance.
1991
Fig. 3. Simplied effect of DR on electricity market prices.
The benets of DR programs are not only for program partici-
pants; some are market-wide. An overall electricity price reduction
is expected eventually because of a more efcient utilization of
the available infrastructure, as in, for example, the reduction of
demand from expensive electricity generating units. Moreover, DR
programs can increase short-term capacity using market-based
programs, which in turn, results in an avoided or deferred capac-
ity costs. The cascaded impact of DR programs includes avoided
or deferred need for distribution and transmission infrastruc-
ture enforcements and upgrades [2,3,9]. All of the avoided or
deferred costs will be reected in the price of electricity for
all electricity consumers (DR programs participants and non-
participants).
Reliability benets can be considered as one of the market-wide
benets because they affect all market participants. Because of their
importance, we have considered reliability benets as one category
by itself. By having a well-designed DR program, participants have
the opportunity to help in reducing the risk of outages. Simultane-
ously and as a consequence, participants are reducing their own risk
of being exposed to forced outages and electricity interruption. On
the other hand, the operator will have more options and resources
to maintain system reliability, thus reducing forced outages and
their consequences [15].
The last category of DR program benets is improving elec-
tricity market performance [16]. DR program participants have
more choices in the market even when retail competition is not
available. Consumers can manage their consumption since they
have the opportunity to affect the market especially with the
market-based programs and dynamic pricing programs. Actually,
this was the prime driver for many utilities to offer DR programs
especially for large consumers [17]. Another important market
improvement is the reduction of price volatility in the spot mar-
ket. Demand responsiveness reduces the ability of main market
players to exercise power in the market [18]. It has been reported
that a small reduction of demand (5%) could have resulted in
a 50% price reduction during the California electricity crisis in
20002001 [19]. This phenomenon is due to the fact that generation
cost increases exponentially near maximum generation capacity. A
small reduction in demand will result in a big reduction in gener-
ation cost and, in turn, a reduction in electricity price, as shown in
Fig. 3.
In this example, the original demand curve is represented by a
vertical line because it is assumed that the system is without DR
programs. DR programs introduce a negative slope on the original
demand curve, leading to a small reduction in demand and a huge
reduction in price. Although some people might argue about envi-
ronmental benets associated with DR programs, those benets are
evident [11]. Environmental benets of DR programs are numerous
and include better land utilization as a result of avoided/deferred
new electricity infrastructure such as generation units and trans-
mission/distribution lines; air and water quality improvement as
a result of efcient use of resources; and reduction of natural
resources depletion [2].
1992 M.H. Albadi, E.F. El-Saadany / Electric Power Systems Research 78 (2008) 19891996
Fig. 4. Classication of DR costs.
3.2. DR costs
Any DR program involves different kind of costs; Fig. 4 shows a
classication of DR programs costs, where both DR program owners
and participants incur initial and running costs [2]. The program
participant might need to install some enabling technologies to
participate in a DR program: smart thermostats, peak load con-
trols, energy management systems, and onsite generation units. A
response plan or strategy needs to be established so that it can
be implemented in case of an event. These initial costs are usually
paid by the participant; however, technical assistance should be
provided by the program.
Participants running costs are those associated with events.
Depending on the response plan, these costs may vary. A reduction
of comfort may results if a customer decides to reset the thermostat,
which results in customer inconvenience that is difcult to quan-
tify. Other event relevant costs are easier to quantify, for instance,
lost business or rescheduling of industrial processes or activities.
If a participating customer decides to use a backup onsite gen-
eration unit, fuel and maintenance costs need to be considered.
The program owner has to take care of initial and running system
wide costs. Most DR programs involve metering and communica-
tion costs as initial costs. Utilities need to install advanced metering
systems to measure, store and, transmit energy usage at required
intervals, e.g., hourly readings for real time pricings. Running costs
of DR programs include administration and management cost of the
program. Moreover, incentive payments are considered as part of
the running costs of IBP. Upgrading the billing system is necessary
before most DR programs are deployed, especially PBP for enabling
the system to deal with time varying cost of electricity.
Another important cost component before deploying any DR
program is educating eligible customers about the potential ben-
ets of the program. Different DR program choices need to be
explained to potential participants and possible demand response
strategies need to be dened. A successful DR program depends
heavily on customer education. Continuous marketing is impor-
tant to attract new participants. Further, a continuous evaluation
Fig. 5. Price elasticity around (po , qo ).
and assessment of DR programs is important to develop a better
approach for reaching the ultimate objectives of the programs [2].
4. DR measurement and simulation
4.1. DR measurement
The ultimate objective of DR programs is to reduce peak demand.
Actual peak demand reduction is used as an indication of how suc-
cessful a DR program is and to compare DR programs in similar
situations. To normalize this indicator, the percentage peak demand
reduction is used. Percentage and actual peak demand reduction
are used to evaluate IBP.
In addition to peak load reduction, the performance of dynamic
pricing programs is measured using demand price elasticity which
represents the sensitivity of customer demand to the price of elec-
tricity. This can be found by calculating the ratio of the percent
change in demand to the percent change in price (E = Q/P)
[3,19,21]. Usually, pricedemand curve of any commodity is
not linear. Therefore, elasticity is linearized around the initial
pricedemand balance (qo , po ); as seen in Fig. 5.
The elasticity of a substitution measures the rate at which the
customer substitutes off-peak electricity consumption for peak
usage in response to a change in the ratio of peak to off-peak
prices [3]. This kind of elasticity is important in TOU and CPP pric-
ing programs. In [22], elasticity is divided into self-elasticity and
cross-elasticity. Self-elasticity measures the demand reduction in
a certain time interval due to the price of that interval. Cross-
elasticity measures the effect of the price of a certain time interval
on electricity consumption during another interval.
Two types of customers are described in [20]: long-range (LR)
and short-range (SR) customers. LR customers maximize long-term
benets by deciding their demands considering all pricing peri-
ods. SR customers set their demand considering the current pricing
period only. However, real world (RW) consumers, described in
[22], consider both current prices and the prices of one step into
the future.
As Fig. 6 suggests, elasticity is used in conjunction with expected
price to modify expected demand; consequently, demand and
prices will be reduced if prices are above the equilibrium point
(po , qo ); see Fig. 5. The equilibrium point is dened as price and
Fig. 6. Effect of price elasticity in price computation.
 M.H. Albadi, E.F. El-Saadany / Electric Power Systems Research 78 (2008) 19891996 1993

Fig. 8. Six-bus test system.

where Vi is the voltage at bus i; is the angle associated with

the voltage at relevant busses; Yij is the element of the Y-bus
admittance matrix;  ij is the angle associated with Yij ; P and Q
are real and reactive power generation at bus i, respectively. PDi
and QDi are real and reactive power demand at bus i, respectively.
2. Generation limits

Fig. 7. Simulation ow chart. Pi min Pi Pi max (5)

Qimin Qi Qimax (6)

demand of the normal case. In this analysis, one equilibrium point
3. Voltage limits
is assumed for each period.
Vi min Vi Vi max i 1, . . . , NL (7)
4.2. Market simulation  
Vi  = const. i 1, . . . , NG (8)
In this analysis, a single-sided uniform market price is consid-
ered in which the supply-side submits bids for supplying power to NL and NG are the number of load buses and generator buses,
the market operator when the market operator has a forecast of respectively. The voltages at load buses are bounded between min-
demand [1]. These bids reect generator cost functions. The oper- imum and maximum values; whereas the voltages at generation
ator announces the expected prices for the next 24 h. These prices buses are kept at constant values.
act as guidelines for customers to respond to in real time, depend- There are two types of market price formulations: the Loca-
ing on their elasticity and expected prices. This DR is very benecial tional Marginal Prices (LMP) and the Uniform Market Price (UMP).
as it can reduce market prices as well as costs. The process used for In an LMP formulation, electricity prices are location dependent.
simulating DR is shown in Fig. 7 [23]. The market price at each bus is represented by the Lagrangian mul-
tiplier i of the real power balance constraint at that bus. In an
4.3. Optimal Power Flow formulation UMP formulation, the market price is the highest value of the bus
incremental cost obtained by solving the above model:
An Optimal Power Flow (OPF) code was developed in a GAMS
 i , i 1, . . . , N (9)
environment to simulate market prices. The objective function of
the OPF is to minimize the total cost of generation for social welfare where  represents the uniform electricity market price; i is the
maximization [1]. incremental cost of generation at bus i, and N is the number of buses
in the system.

NG
J= Ci (Pi ) (1)
i=1
4.4. Simulation results

Ci (Pi ) = aPi 2 + bPi + ci
where J is the total generation costs; Ci (Pi ) is the cost function of
generator i; Pi is the power output of generator i, and NG is the
number of generators. Generators are assumed to be bidding their
true cost of generation. The minimization objective function has
the following constraints:
(2) A six-bus system, presented in [1], was used for this simulation.
The system is presented in Fig. 8, and Tables 1 and 2 show the system
data. The generation cost function ($/MWh) of the two generation
units is represented by the quadratic functions below.
Table 1
Line data

1. Power balance equations Line R (p.u.) X (p.u.) B/2 (p.u.)

   14 0.0662 0.1804 0.003
Pi PDi = Vi  Vj  Yij cos(i,j + j i ) (3) 16 0.0945 0.2987 0.005
j
23 0.0210 0.1097 0.004
25 0.0824 0.2732 0.004
   34 0.1070 0.3185 0.005
Qi QDi = Vi  Vj  Yij sin(i,j + j i ) (4) 46 0.0639 0.1792 0.001
j 56 0.0340 0.0980 0.004
1994 M.H. Albadi, E.F. El-Saadany / Electric Power Systems Research 78 (2008) 19891996

Table 2
Bus data with Sbase = 100 MV A

# Demand (p.u.) Generation limits (p.u.)

PD QD Pmax Pmin Qmax Qmin

1 0.73125 0.1950 5.0 1.0 3.0 0.2

2 0.92625 0.2925 2.5 0.5 1.5 0.2
3 0.78000 0.3900 0.0 0.0 0.0 0.0
4 1.12125 0.3120 0.0 0.0 0.0 0.0
5 1.26750 0.34125 0.0 0.0 0.0 0.0
6 0.67375 0.24375 0.0 0.0 0.0 0.0

Fig. 11. A comparison between base-case, contingency and prices when loads have
0.1 elasticity.

response to the day-ahead published prices, the very high expected

prices of the two peaks were actually reduced even below the
base-case prices. However, this will not happen in practice because
elasticity is not constant, and it has a lower value during peak peri-
ods.

4.5. DR experiences

Fig. 9. Base-case simulation results.
C1 = P12 + 8.5P1 + 5
C2 = 3.4P22 + 25.5P2 + 9
A Load Scaling Factor (LSF) was used to represent load varia-
tion. Simulation results of the base case are shown in Fig. 9. The
equilibrium point during each interval is the demand and the price
of the base-case scenario. As shown in Fig. 9, 12 price intervals
were considered. The gure shows clearly that electricity prices
are sensitive to increase in demand. A contingency in the system
was introduced by removing line 46 from the network data. Obvi-
ously, this contingency causes a large increase in market price, with
inelastic demand, especially during peak periods (Fig. 10). These
high prices will be announced 24 h ahead. Note that we assume
that the operator is aware that such a contingency will happen the
next day and will last for the whole day.
To illustrate the effect of demand elasticity, loads at all busses
are assumed to have the same constant self-elasticity of 0.1, which
means that a 100% change in price will reduce the load by 10%. All
cross-elasticities are neglected, which means that the customers
are of the SR type. After calculating the change in LSF, the OPF model
is solved, and the results are shown in Fig. 11.
It is notable that real prices were reduced below contingency
prices. Moreover, due to the assumption that customers react in
Fig. 10. A comparison between base-case and contingency case prices.
Three types of DR quantization studies have been distinguished
[2]: illustrative studies, integrated resource planning studies, and
program evaluation studies. It has been shown that program eval-
(10)
uation studies revealed much lower benets than the other two
(11) studies. Illustrative studies assume high penetration rates and long-
term sustained benets. Similarly, integrated recourse planning
studies consider long-term benets. On the other hand, program
evaluation studies do not consider long-term benets and suffer
from low penetration rates. Many utilities in North America and
around the globe have experiences with IBP. As an example, NYISO
IBP paid out US$ 7.2 million in incentives to more than 14,000 pro-
gram participants to release 700 MW peak capacity in the summer
of 2003 [12]. The load curtailment programs were estimated pro-
viding reliability benets of more than US$ 50 million on August
15, 2003 [12]. In general, it was reported that the benets of these
programs exceeded the cost by a factor of 7:1 [12]. TOU pricing is
the basic PBP and easiest to implement. Electricite De France (EDF)
operates what is probably the most successful example of a TOU
pricing program. This program was applied to large industrial cus-
tomers in 1956 and introduced to residential customers in 1965.
Currently, it is estimated that one third of its customers are on TOU
pricing [12]. In 1993, EDF introduced a CPP pricing program called
Tempo in which the year is divided into three types of days: Tempo
Blue, Tempo White, and Tempo Red. 300 days of the years are Tempo
Blue during which time electricity is cheaper than the normal TOU
prices. Tempo White days are 43 and they are at a slightly higher
rates compared to that of normal TOU. Tempo Red days are only
21 and they are the most expensive. Customers can know the color
of the next day by several means. TOU pricing was implemented
by many utilities in North America. An experiment implemented
in Pennsylvania revealed an average elasticity of substitution of
0.14 [12]. This means that a 100% price increase will correspond
to 14% reduction in demand. Another experiment, in Florida by the
Gulf Power Company, used TOU pricing. Customers were provided
with smart thermostats that automatically adjusted the tempera-
ture and other loads depending on a price signal. In this program,
normal TOU prices were applied 99% of all hours in the year. In the
remaining 1% of the hours, the utility had the option of charging
a critical peak pricing, more than the normal peak period price.
This program resulted in 42% peak demand reduction during criti-
cal peak periods [12]. A comprehensive survey of utility experience
 M.H. Albadi, E.F. El-Saadany / Electric Power Systems Research 78 (2008) 19891996
with RTP was presented in [17]. This survey covered 43 voluntary
real time pricing programs offered in 2003. It has been reported that
the most common utility motivation behind these programs was
customer satisfaction by providing opportunities for bill savings.
Encouraging peak demand reduction and load growth comes after
the prime motivation. Complying with new regulations was also
mentioned as a motivation. It has been reported that penetration
rates were low in most programs. In some programs, penetration
levels were dropping even lower. The problem of low program par-
ticipation was attributed to poor marketing and limited technical
assistance provided to help participants managing price volatility.
Most RTP participants were large industrial customers and some
large institutional ones. This survey concluded that big portions of
required information about price responsiveness were not avail-
able in most programs. In addition, some RTP participants are not
price responsive at all. Price responsive customers generally employ
on-site generation or simple strategies like rescheduling. Some of
these participants were found to be very sensitive to prices as low as
US$ 0.20/kWh. The programs under study were reported to achieve
1233% aggregate load reduction across a wide range of prices. Only
one program was able to generate a load reduction of more than
1% of the utility system peak [17]. Another case study of the Nia-
gara Mohawk RTP program in New York found that the average
substitution elasticity was 0.14 [13].
5. Conclusions
DR changes end-use customer electricity consumption patterns
from those customers normal patterns in response to changes
in the price of electricity over time, or to incentive payments
designed to induce lower electricity use at times of high whole-
sale market prices, or when system reliability is jeopardized. DR
program benets cover all electricity consumers. DR programs can
reduce electricity prices, improve system reliability, and reduce
price volatility. To employ DR programs, both participants and pro-
gram owners incur initial and running costs. The performance of DR
programs is measured by peak load reduction and demand elastic-
ity. Although illustrative studies and integrated resource planning
studies project more benets from DR programs, program evalua-
tion studies prove the substantial benets of these programs. The
above case study simulating the effect of elasticity in electricity
prices demonstrates the effect of DG programs in cases of system
contingency.
Appendix A. List of symbols
Ci (Pi ) the cost function of generator i
CPP Critical Peak Pricing
DR Demand Response
ED-CPP Extreme Day Critical Peak Pricing
EDF Electricite De France
EDP Extreme Day Pricing
IBP Incentive-Based Programs
J the total generation costs
LMP Locational Marginal Prices
LR long-range customers
LSF Load Scaling Factor
N the number of buses in the system
NG the number of generators
NYISO New York Independent System Operator
OPF Optimal Power Flow
Pi the real power output of generator i
PDi real power demand at bus i
1995
Qi The reactive power output of generator i
QDi reactive power demand at bus i
RTP Real Time Pricing
RW real world customers
SR short-range customers
TOU Time of Use
UMP Uniform Market Price
V the bus voltage
Yij The element of the Y-bus admittance matrix
Greek symbols
the angle associated with the voltage at relevant busses
i the incremental cost of generation at bus i
 ij the angle associated with Yij
 the uniform electricity market price
References
[1] K. Bhattacharya, M. Bollen, J. Daalder, Operation of Restructured Power System,
Kluwer Academic Publishers, London, 2001.
[2] US Department of Energy, Benets of Demand Response in Electricity Mar-
kets and Recommendations for Achieving them, Report to the United States
Congress, February 2006, available online: http://eetd.lbl.gov.
[3] International Energy Agency, The Power to ChooseDemand Response in Lib-
eralized Electricity Markets, OECD, Paris, 2003.
[4] K. Herter, P. McAuliffe, A. Rosenfeld, An exploratory analysis of California res-
idential customer response to critical peak pricing of electricity, Energy 32
(2007) 2534.
[5] M. Piette, O. Sezgen, D. Watson, N. Motegi, C. Shockman, Development and Eval-
uation of Fully Automated Demand Response in Large Facilities, Prepared For
California Energy Commission, Public Interest Energy Research (PIER) Program,
March 30, 2004 (LBNL-55085).
[6] S. Valero, M. Ortiz, C. Senabre, C. Alvarez, F. Franco, A. Gabaldon, Methods for
customer and demand response policies selection in new electricity markets,
IET Gen. Transm. Distrib. 1 (2007) 104110.
[7] O. Sezgen, C.A. Goldman, P. Krishnarao, Option value of electricity demand
response, Energy 32 (2007) 108119.
[8] M.H. Albadi, E.F. El-Saadany, Demand response in electricity markets: an
overview, in: IEEE PES GM, Montreal, 2007, pp. 15.
[9] Y. Tan, D. Kirschen, Classication of Control for Demand-side Participation,
University of Manchester, 29 March 2007.
[10] P. Jazayeri, A. Schellenberg, W.D. Rosehard, J. Doudna, S. Widergren, D.
Lawrence, J. Mickey, S. Jones, A survey of load control programs for price and
system stability, IEEE Trans. Power Syst. 20 (3) (2005) 15041509.
[11] C. Gellings, J. Chamberlin, Demand Side Management: Concepts and Methods,
The Fairmont Press Inc., US, 1988.
[12] Charles River Associates, Primer on Demand-Side Management with an Empha-
sis on Price-Responsive Programs, Report prepared for The World Bank,
Washington, DC, CRA No. D06090, available online: http://www.worldbank.
org.
[13] E. Bloustein, School of Planning and Public Policy, Assessment of Customer
Response to Real Time Pricing, RutgersThe State University of New Jersey,
June 30, 2005, available online: http://www.policy.rutgers.edu.
[14] D.S. Kirschen, Demand-side view of electricity markets, IEEE Trans. Power Syst.
18 (2003) 520527.
[15] L. Goel, W. Qiuwei, W. Peng, Reliability enhancement of a deregulated power
system considering demand response, in: IEEE PES GM, 2006, pp. 16.
[16] K. Spees, L. Lave, Demand response and electricity market efciency, Electricity
J. 20 (2007) 6985.
[17] G. Barbose, C. Goldman, C. Neenan, A Survey of Utility Experience with Real
Time Pricing, Lawrence Berkeley National Laboratory Report No. LBNL-54238,
December 2004.
[18] D. Caves, K. Eakin, A. Faruqui, Mitigating price spikes in wholesale markets
through market-based pricing in retail markets, Electricity J. 13 (2000) 1323.
[19] S. Braithwait, K. Eakin, The role of demand response in electric power market
design, Laurits R. Christensen Associates, Prepared for Edison Electric Institute,
Madison, October 2002.
[20] A.K. David, Y.Z. Li, Consumer rationality assumptions in the real time pricing of
electricity, IEE Proc. Gen. Trans. Dist. 139 (1992) 315322.
[21] M.G. Lijesen, The real-time price elasticity of electricity, Energy Econ. 29 (2007)
249258.
[22] D.S. Kirschen, G. Strbac, P. Cumperayot, D. Mendes, Factoring the elasticity of
demand in electricity prices, IEEE Trans. Power Syst. 15 (2000) 612617.
[23] E.C. Banda, L.A. Tuan, Modeling of demand response in electricity markets:
effects of price elasticity, in: Power and Energy Systems, Florida, January 2007.
Mohamed H. Albadi was born in Sohar, Oman, in 1976. He received the B.Sc. degree in
electrical and computer engineering from Sultan Qaboos University, Muscat, Oman
in 2000 and M.Sc. degree in electrical engineering from Faculty of Engineering,
1996 M.H. Albadi, E.F. El-Saadany / Electric Power Systems Research 78 (2008) 19891996

Aachen University of Technology, Aachen, Germany in 2003. He is currently pur-
suing the Ph.D. degree in the Department of Electrical and Computer Engineering,
University of Waterloo, Waterloo, ON, Canada. His research interests are energy
management, distributed generation, and power quality.
Ehab F. El-Saadany was born in Cairo, Egypt, in 1964. He received the B.Sc. and M.Sc.
degrees in electrical engineering from Ain Shams University, Cairo, Egypt, in 1986
and 1990, respectively, and the Ph.D. degree in electrical engineering from the Uni-
versity of Waterloo, Waterloo, ON, Canada, in 1998. Currently, he is an associate
professor in the Department of Electrical and Computer Engineering, University
of Waterloo. His research interests are distribution system control and operation,
power quality, distributed generation, power electronics, digital signal processing
applications to power systems, and Mechatronics.